Papers for

text-to-speech engineers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Community-aware sampling improves text-to-speech training efficiency

Structure Before Sampling: Community-Aware Core-Set Selection for Data-Efficient Text-to-Speech

Abstract: Text-to-speech (TTS) corpora are costly to record, yet many utterances add little new phonetic information. Core-set selection reduces this cost by choosing a small training subset under a fixed audio-duration budget. We represent a corpus as a phonotactic graph that links each utterance to its most phonemically similar ones, and we first test whether this graph has structure. In Bangla and English corpora, its clustering is 199 and 56 times that of a size-matched random graph, and its modularity is more than twice that of a degree-preserving random graph. We then propose Community Representative, a selector that samples across graph communities and spreads its choices within each one, starting from utterances rich in rare phonemes. At every budget and in both languages, it covers more rare phoneme bigrams than random and entropy-based selection, and this lead holds on held-out utterances. TTS models trained on its 20% core-sets have a significantly lower character error rate (CER) than models trained on equal-duration random or entropy-based subsets in both languages. When all models train for the same number of epochs, the Bangla core-set model also outperforms full-corpus training (3.93% vs. 4.47% CER) with 4.5x less training time.

Mon 21 SeptComputation and Language
The gist
Recording speech for computer voices is expensive, and many recorded sentences don’t add new sound information. The authors found that grouping similar sentences by their sounds reveals strong clusters. They created a method that picks sentences from different clusters and those with rare sounds to cover more sound variety efficiently. Training text-to-speech models on these smartly chosen sentences leads to better performance using much less data and time.
Open → 2609.24275v1

Neural grapheme to phoneme method improves reading in unsegmented languages

Dictionary-Constrained Grapheme-to-Phoneme for Unsegmented Languages from LLM-Annotated Data

Abstract: Grapheme-to-phoneme (G2P) conversion turns raw text into its phonemic form and is an essential part of both text-to-speech (TTS) and automatic speech recognition (ASR) systems. It is required to be fast, stable and context-aware. For unsegmented languages such as Japanese, G2P additionally couples word segmentation with highly context-dependent polyphone disambiguation, and the scarcity of accurately annotated data remains a bottleneck. In this paper, we present a context-aware neural G2P method that scores paths of a discriminative conditional random field (CRF) over a word lattice constructed from dictionaries. To tackle data scarcity, we utilize large language models (LLMs) to generate more than 2 million sentences. Experimental results demonstrate that our method strongly outperforms conventional morphological analyzer-based methods and neural sequence models. On the Joyo-Kanji-Yomi benchmark, our method reaches 99.62% target word reading accuracy, 0.32% target word phoneme error rate (PER) and 0.14% sentence PER.

Thu 17 SeptComputation and Language
The gist
Languages like Japanese don’t clearly separate words in text, making it hard for computers to convert written characters into their correct sounds. The authors created a smart computer method that uses dictionary guesses and scores to help figure out pronunciations accurately. They also used large language AI to create millions of example sentences, which helped train their system better. Their new method works much better than older approaches, getting almost perfect reading accuracy on a common Japanese benchmark.
Open → 2609.19805v1